IEEE Trans Neural Syst Rehabil Eng. 2018 Aug;26(8):1596-1603. doi: 10.1109/TNSRE.2018.2854605. Epub 2018 Jul 9.
Lower extremity powered exoskeletons (LEPEs) allow people with spinal cord injury (SCI) to stand and walk. However, the majority of LEPEs walk slowly and users can become fatigued from overuse of forearm crutches, suggesting LEPE design can be enhanced. Virtual prototyping is a cost-effective way of improving design; therefore, this research developed and validated two models that simulate walking with the Bionik Laboratories' ARKE exoskeleton attached to a human musculoskeletal model. The first model was driven by kinematic data from 30 able-bodied participants walking at realistic slow walking speeds (0.2-0.8 m/s) and accurately predicted ground reaction forces (GRF) for all speeds. The second model added upper limb crutches and was driven by 3-D-marker data from five SCI participants walking with ARKE. Vertical GRF had the strongest correlations (>0.90) and root-mean-square error (RMSE) and mediolateral center of pressure trajectory had the weakest (<0.35), for both models. Strong correlations and small RMSE between predicted and measured GRFs support the use of these models for optimizing LEPE joint mechanics and improving LEPE design.
下肢助力外骨骼（LEPE）允许脊髓损伤（SCI）患者站立和行走。然而，大多数 LEPE 行走缓慢，并且由于过度使用前臂拐杖，用户可能会感到疲劳，这表明 LEPE 的设计可以得到改进。虚拟原型设计是一种具有成本效益的改进设计的方法；因此，本研究开发并验证了两个模型，这些模型模拟了 Bionik Laboratories' ARKE 外骨骼附着在人体肌肉骨骼模型上的行走。第一个模型由 30 名健康参与者以现实的慢走速度（0.2-0.8 m/s）行走的运动学数据驱动，并准确预测了所有速度下的地面反作用力（GRF）。第二个模型添加了上肢拐杖，并由 5 名使用 ARKE 行走的 SCI 参与者的 3D 标记数据驱动。对于这两个模型，垂直 GRF 的相关性最强（>0.90），均方根误差（RMSE）和横向中心压力轨迹的相关性最弱（<0.35）。预测和测量的 GRF 之间的强相关性和较小的 RMSE 支持使用这些模型来优化 LEPE 关节力学和改进 LEPE 设计。
